LSHash icon indicating copy to clipboard operation
LSHash copied to clipboard

Is it possible to query results based on threshold.

Open tom-riddle opened this issue 6 years ago • 3 comments

Many LSH implementations use Jaccard similarity to return matching result above a certain threshold say 80% match. Is possible to implement the same in this library.

tom-riddle avatar Jul 28 '18 16:07 tom-riddle

Same thoughts. And I think zipping arrays into lower dimensions and using smaller input_dim may help, as smaller dimensions increases the probability of collision, and therefore similar vectors are more likely fall into a same slot.

Riroaki avatar Jul 17 '19 04:07 Riroaki

Same thoughts. And I think zipping arrays into lower dimensions and using smaller input_dim may help, as smaller dimensions increases the probability of collision, and therefore similar vectors are more likely fall into a same slot.

Sorry, I mean hash_size..

Riroaki avatar Jul 17 '19 06:07 Riroaki

Many LSH implementations use Jaccard similarity to return matching result above a certain threshold say 80% match. Is possible to implement the same in this library.

How to use threshold with this LSH implementation can you help me I have problem with this issue

AmeerahAlshahrani avatar Jan 17 '23 11:01 AmeerahAlshahrani